Home page

Main topics
News Headlines

Frequently asked Questions
For Policy Makers

Research evidence
Misleading claims
Helmet laws

Search Engine

New Zealand
Other countries

Full index

Policy statement


Weak study claims helmet laws lead to children quiting cycling

According to a study by Carpenter and Stehr, helmet laws encourage kids to give up cycling in the US. Even then a reduction of only 8 deaths per year in US states with helmet laws can be associated with such laws. It takes 81,000 fewer riders for every life saved!

However, the study is weak because it depended on self-reported data about both helmet wearing and bicycle use. Some smaller nitpicks: C&S provide no information on the amount of missing data, except to say that they have omitted children with missing relevant data. It would be good practice to say what proportion of the total survey population are thus omitted - if it's really high it leaves room for plenty of inaccuracy. They also provide rather little in the way of tabulations of the input data - providing these is good practice so that the reader can check if the results look credible.

Carpenter and Stehr associated helmet laws with an increase in "always or almost always wears helmet" when cycling of 34.9% (on parental report) or 9.7% (self-report), a 19% reduction in fatalities (which should be accurate as data gets, because it's based on death registration data) and a 3.8% (parental report) or 4.7% (self-report) reduction in the number who had ever ridden a bike in the last 12 months.

A further problem is that the original data weren't designed to assess the actual amount of cycling at all, there is no data on how often or how far any of these children rode. C&S do make this point, their note (18 in Jan, 21 in March) reads: "Although we estimate that bicycling participation fell by about 5 percent, it is likely that overall bicycling miles travelled fell even more. Unfortunately, neither the YRBSS nor the BRFSS asked consistent questions about bicycling intensity over the full sample period. As noted previously, however, the YRBSS did ask about the number of instances of bike riding from 1991 to 1997. We estimated equation (1) on this outcome (using the midpoints of the ranges and coding the top category as 50 instances) and found that helmet laws reduced bicycling among high school youths age 15 and under by 2.34 instances, or about 11 percent relative to the pre-reform mean of 21.32 instances. This suggests that the true overall reduction in bicycling miles travelled – and thus exposure to potential bicycling accidents – is larger than our bicycling participation estimates in Tables 3 and 4. These estimates are of course based on fewer state changes and as such are less precise than the results for bicycle riding, which we observe over the entire sample period from 1991 to 2005 in the YRBSS."  It would strike me as perfectly credible on this basis to suggest that risk of death per mile was higher with helmet laws. That too would be a very weak conclusion. Overall I'd summarize the study as a brave attempt to extract some interesting suggestions from data that aren't strong enough to support any real conclusions one way or another.

Sun 10 Oct 2010